<joanie> agenda: this
<joanie> agenda: be done
<scribe> scribe: janina
jd: Two specs depend on accname, so we need to get accname out so those other two can move
mc: accname needs to be as mature as the other specs, i.e. cr if they're cr
jd: Noting that the dependencies
aren't just for the aam, but on the actual algo
... This got me to put a list of tasks together than can help
us as a group get this out
<joanie> 1. Port testable statements from ARIA 1.0 to wiki (done by Joanie)
jd: Done!
<joanie> https://www.w3.org/wiki/AccName_1.1_Testable_Statements
jd: Mostly tests Michael exported
from 1.0
... Expected results have changed, though
... test names point to 1.0
<joanie> https://github.com/WhatSock/w3c-alternative-text-computation
jd: uses a tool that presents
expected results environment by environment
... Have ported
<joanie> 2. Port any missing ARIA 1.0 description related tests to wiki (will be done by Joanie + Michael)
bg: just a js function
jd: Joanie and Michael will do
<joanie> 3. Convert wiki tests to WPT and add to WPT repo (being done by Joanie)
jd: Essentially done--waiting on
a pull request
... In the works--expecting from Shane soon
<joanie> 4. Gathering test results for all platforms (Joanie will coordinate and provide Linux and Mac results, Jon Gunderson for IA2, UIA? Jon or MS)
jd: Will coordinate all
... JG, can you do IA2?
jg: Probably also UI Automate
jd: OK, tentavily for UIA
<joanie> 5. Validate test cases using Bryan's tool (hopefully can assign to bryan)
<joanie> https://www.w3.org/wiki/AccName_1.1_Testable_Statements
jd: Level 3 headings -- if statements , please port between if and then
bg: Actually put those conditions in?
jd: Yes, so can click on a button and get the expected results
bg; Will look
jd: Noting most are simple, and
hoping they're not wrong
... We're hoping to learn the expected results are actually
correct
... If we can't predict without pulling out tools, maybe we're
too complex. After 1.1, we'll return to discuss this issue
bg: noting certain areas of spec
don't match realworld expectations
... part of my reason for working on a tool
<joanie> 6. Identify missing test cases, i.e. something in algorithm for which we lack coverage (assignee?). RELATED: LET'S BE SIMPLE in our cases
jd: Not convinced we have all
needed tests
... was looking at a span that yielded different results
everywhere
<joanie> <span role="presentation" aria-label="Eli">
<joanie> <span aria-label="Garaventa">Zambino</span>
<joanie> </span>
jd: above 3 lines example of
missing test
... table would be another
... we also want to make it easy for devs to file bugs
mk Thinking of situations which required html knowledge in addition to accname
mk: i.e. sometimes the answer is
in the html spec
... should we be referring to relevant parts of other
specs?
jd: let's make that a 1.2
question
... Need everyone's help on identifying missing test cases.
Please add!
... Don't expect cases added to wiki ...
<joanie> 7. Triage issues filed against AccName, i.e. 1.1 or 1.2. For 1.1, obtain proposed changes to algorithm and get pull requests (assignee?)
jd: i.e. 1.1 or 1.2?
<joanie> https://github.com/w3c/accname/issues
jd: Need to triage that list
mk: Can take a first pass,
certainly
... Understand if relatively easy, propose a solution for
1.1
jd: Correct
mk: A bit concerned of timeline
..... Next week a short week
... Believe could do week following next
jd: Noting also that we need to figure out when we can go CR, because that sets a clock
jd: basically, this is what mk is
looking at
... any other
jd: BG will do the verification
part of tests on wiki ... porting to his tool
... I can followup with sanity checking
[discussion of accname that could be 1.2]
mk: wondering if there other things in the spec not clearly spelled out that bg could document?
bg: 3 areas ...
... browsers are filling gaps that are unclear
... white space is 1
... ties to block vs inline
... also multiple ids
... two levels to issue between block level and inline
elements
... css styling related
... then also css pseudo elements
... label elements another -- when associated with same form
field -- aminor thing
mk: will be mindful as
triaging
... inclined to leave css out
... believe that's 1.2, though
jd: agreed
... No meaty issues at this point! Those are 1.2
bg: we need to agree on what should be happening
jd: yes
jd: Who's consuming accname?
<joanie> If assistive technologies utilize the accessibility tree to convey this information, then these browser improvements will automatically increase accessibility within these assistive technologies.
jd: ats consume info from tree to
convey the info
... Believe we want browsers to be the only consumers of
accname
bg: agree the goal, but noting it's not what's happening now
mk: can't blame ie11, though
jd: Expressing thanks all around--lots of awesome work, especially bg's tool!
This is scribe.perl Revision: 1.152 of Date: 2017/02/06 11:04:15 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Default Present: Joanmarie_Diggs, Irfan_Ali, jamesn, jongund, mat_king, Stefan, janina, matt_king Present: Joanmarie_Diggs Irfan_Ali jamesn jongund mat_king Stefan janina matt_king Regrets: Irfan_Ali Found Scribe: janina Inferring ScribeNick: janina WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth Found Date: 15 Feb 2018 People with action items: WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option. WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]